On Mon, 19 Oct 2015, Paolo Chirico wrote: > Hi, > when I run Kalman filter, some matrices are generated in the session > (and in the icons window). > I save the session, but when I open that session, all matrices are > missing. > Why? Could you tell us what version of gretl you're running, and give us an idea of what Kalman functions you're running? Thanks. Allin Cottrell _______________________________________________ Gretl-users mailing list Gretl-users(a)lists.wfu.edu http://lists.wfu.edu/mailman/listinfo/gretl-users

nulldata 50 # may be, the key moment

On Mon, 19 Oct 2015, Paolo Chirico wrote: > Il 19/10/2015 13:25, Allin Cottrell ha scritto: >> On Mon, 19 Oct 2015, Paolo Chirico wrote: >> >>> Hi, >>> when I run Kalman filter, some matrices are generated in the session (and >>> in the icons window). >>> I save the session, but when I open that session, all matrices are >>> missing. >>> Why? >> >> Could you tell us what version of gretl you're running, and give us an idea >> of what Kalman functions you're running? Thanks. >> > I'm using the last version and the matices are generated by kfilter function > (kfilter(&E,&V,&S,&P,&K)). > I guess the problem concerns the updating, but It's only a idea of mine. I think there may be a misunderstanding here. I open and run the following trivial Kalman script in gretl: nulldata 50 setobs 1 1 --special scalar s1 = 1 scalar s2 = 1 series y = normal() kalman obsy y obsymat 1 statemat 1 statevar s2 obsvar s1 end kalman --diffuse matrix E V S MSE G kfilter(&E, &V, &S, &MSE, &G) # optional: take a look print E V S I then save the session, either by going to /File/Session files/Save session, or by right-click in the "Icon view" window. I call the file kaltest.gretl. I then exit from gretl, restart the program, and open kaltest.gretl (e.g. under the /File/Session files menu). The matrices E, V, S, MSE and G are shown in the Icon view window. Where does your experience differ? Allin Cottrell _______________________________________________ Gretl-users mailing list Gretl-users(a)lists.wfu.edu http://lists.wfu.edu/mailman/listinfo/gretl-users

On Mon, 19 Oct 2015, oleg_komashko(a)ukr.net wrote: > Dear Allin, > In April I lost a nice set of simulations > not using nulldata > > Now I made mmm = mnormal(4,4) > without any data and it was saved OK > It is difficult to say what is Paolo's last version: > last git, last release or last version on some Linux repo > >> nulldata 50 # may be, the key moment Hmm, I just ran a variant of my test script in http://lists.wfu.edu/pipermail/gretl-users/2015-October/011242.html This time I didn't use "nulldata" (and so didn't have a dataset in place) but instead defined the y variable as a matrix: matrix y = mnormal(50,1) This didn't make a difference to the matrix-saving (which still worked OK). Perhaps having or not having a dataset in place made a difference at one time, though I don't remember that being reported as a bug. Allin _______________________________________________ Gretl-users mailing list Gretl-users(a)lists.wfu.edu http://lists.wfu.edu/mailman/listinfo/gretl-users